Graded Elevation of C-Jun in Schwann Cells in Vivo: Gene Dosage Determines Effects on Development, Remyelination, Tumorigenesis, and Hypomyelination

The Journal of Neuroscience, December 13, 2017 • 37(50):12297–12313 • 12297

Development/Plasticity/Repair Graded Elevation of c-Jun in Schwann Cells In Vivo: Gene Dosage Determines Effects on Development, Remyelination, Tumorigenesis, and Hypomyelination

Shaline V. Fazal,1* XJose A. Gomez-Sanchez,1* Laura J. Wagstaff,1 Nicolo Musner,2 XGeorg Otto,3 Martin Janz,4 Rhona Mirsky,1 and Kristján R. Jessen1 1Department of Cell and Developmental Biology, University College London, London WC1E 6BT, United Kingdom, 2Enzo Life Sciences, 4415 Lausen, Switzerland, 3University College London Great Ormond Street Institute of Child Health, London WC1N1EH, United Kingdom, and 4Max Delbru¨ck Center for Molecular Medicine and Charite´, University Hospital Berlin, Campus Benjamin Franklin, 13092 Berlin, Germany

Schwann cell c-Jun is implicated in adaptive and maladaptive functions in peripheral nerves. In injured nerves, this transcription factor promotes the repair Schwann cell phenotype and regeneration and promotes Schwann-cell-mediated neurotrophic support in models of peripheral neuropathies. However, c-Jun is associated with tumor formation in some systems, potentially suppresses myelin genes, and has been implicated in demyelinating neuropathies. To clarify these issues and to determine how c-Jun levels determine its function, we have generated c-Jun OE/ϩ and c-Jun OE/OE mice with graded expression of c-Jun in Schwann cells and examined these lines during development, in adulthood, and after injury using RNA sequencing analysis, quantitative electron microscopic morphometry, Western blotting, and functional tests. Schwann cells are remarkably tolerant of elevated c-Jun because the nerves of c-Jun OE/ϩ mice, in which c-Jun is elevated ϳ6-fold, are normal with the exception of modestly reduced myelin thickness. The stronger elevation of c-Jun in c-Jun OE/OE mice is, however, sufficient to induce significant hypomyelination pathology, implicating c-Jun as a potential player in demyelinating neuropathies. The tumor suppressor P19 ARF is strongly activated in the nerves of these mice and, even in aged c-Jun OE/OE mice, there is no evidence of tumors. This is consistent with the fact that tumors do not form in injured nerves, although they contain proliferating Schwann cells with strikingly elevated c-Jun. Furthermore, in crushed nerves of c-Jun OE/ϩ mice, where c-Jun levels are overexpressed sufficiently to accelerate axonal regeneration, myelination and function are restored after injury. Key words: c-Jun; myelin; PNS; regeneration; Schwann; tumorigenesis

Significance Statement In injured and diseased nerves, the transcription factor c-Jun in Schwann cells is elevated and variously implicated in controlling beneficial or adverse functions, including trophic Schwann cell support for neurons, promotion of regeneration, tumorigenesis, and suppres- sionofmyelination.Toanalyzethefunctionsofc-Jun,wehaveusedtransgenicmicewithgradedelevationofSchwanncellc-Jun.Weshow that high c-Jun elevation is a potential pathogenic mechanism because it inhibits myelination. Conversely, we did not find a link between c-Jun elevation and tumorigenesis. Modest c-Jun elevation, which is beneficial for regeneration, is well tolerated during Schwann cell development and in the adult and is compatible with restoration of myelination and nerve function after injury.

Introduction this transcription factor is a global amplifier of the repair Schwann Schwann cell c-Jun has been implicated both in adaptive and cell phenotype and promotes regeneration and, in models of pe- maladaptive functions in peripheral nerves. In injured nerves, ripheral neuropathies, Schwann cell c-Jun supports axonal sur-

*S.V.F. and J.A.G.-S. contributed equally to this work. Received April 12, 2017; revised Sept. 22, 2017; accepted Oct. 8, 2017. Correspondence should be addressed to either Kristjan R. Jessen or Rhona Mirsky, Department of Cell and Author contributions: R.M. and K.R.J. designed research; S.V.F., J.A.G.-S., L.J.W., and N.M. performed research; Developmental Biology, University College London, Gower Street, London WC1E 6BT, United Kingdom, E-mail: M.J.contributedunpublishedreagents/analytictools;S.V.F.,J.A.G.-S.,G.O.,andK.R.J.analyzeddata;R.M.andK.R.J. [email protected] or [email protected]. wrote the paper. DOI:10.1523/JNEUROSCI.0986-17.2017 This work was supported by the Wellcome Trust (Programme Grant 074665 to K.R.J. and R.M.), the Medical Copyright © 2017 Fazal, Gomez-Sanchez et al. Research Council (Project Grant G0600967 to K.R.J. and R.M.), and the European Community (Grant HEALTH-F2- This is an open-access article distributed under the terms of the Creative Commons Attribution License 2008-201535 from FP7/2007-3013). We thank M. Turmaine for providing help and advice for electron microscopy. Creative Commons Attribution 4.0 International, which permits unrestricted use, distribution and reproduction in The authors declare no competing financial interests. any medium provided that the original work is properly attributed. 12298 • J. Neurosci., December 13, 2017 • 37(50):12297–12313 Fazal, Gomez-Sanchez et al. • Enforced Expression of c-Jun in Schwann Cells vival, trophic factor expression, and sensory–motor function Materials and Methods (Arthur-Farraj et al., 2012, 2017; Hantke et al., 2014; Klein et al., Transgenic mice 2014; Jessen and Mirsky, 2016). Reduced c-Jun levels in Schwann Animal experiments conformed to UK Home Office guidelines under the cells are also implicated in failure of regeneration due to aging supervision of University College London (UCL) Biological Services. To and long-term denervation (Painter et al., 2014; Jessen and Mir- generate mice that overexpress c-Jun selectively in Schwann cells, female sky, 2016). The c-Jun pathway is therefore of interest for the R26c-Junstopf mice, generated in the laboratory of Klaus Rajewsky, which carry a lox-P flanked STOP cassette in front of a CAG promoter-driven development of a pharmacology for nerve repair. Conversely, ϩ Ϫ c-Jun cDNA in the ROSA26 locus, were crossed with male P0Cre / c-Jun is associated with tumor formation in some systems and mice (Feltri et al., 1999). This generated P0-Cre ϩ/ Ϫ;R26c-Junstopf f/ϩ potentially suppresses myelin genes (Eferl and Wagner, 2003; mice, which we refer to as c-Jun OE/ϩ mice. These male mice were Parkinson et al., 2008). Based on this, it has been characterized as back-crossed with female R26c-Junstopf f/f mice to generate P0-Cre ϩ; a negative regulator of myelination and implicated in demy- R26c-Junstopf f/f mice, referred to as c-Jun OE/OE mice. P0-Cre Ϫ / Ϫ lit- elinating neuropathies (Jessen and Mirsky, 2008). In the present termates were used as controls. Mice of either sex and on the C57BL/6 work, we have generated and analyzed mouse lines with graded background were used in the experiments. expression of c-Jun in Schwann cells to clarify these issues and to determine how c-Jun levels determine its function. Genotyping DNA for genotyping was extracted from ear or tail samples using the c-Jun is present at low levels in Schwann cells of uninjured HotSHot method (Truett et al., 2000). Primers for genotyping the R26c- nerves, but is rapidly elevated 80- to 100-fold after nerve cut (De Junstopf transgene were 5Ј-TGGCACAGCTTAAGCAGAAA-3Ј and 5Ј- Felipe and Hunt, 1994; Shy et al., 1996; Parkinson et al., 2008;J. GCAATATGGTGGAAAATAAC-3Ј (270bp). The primers for the Rosa26 Gomez-Sanchez, K.R. Jessen, and R. Mirsky, unpublished data). wild-type locus were 5ЈGGAGTGTTGCAATACCTTTCTGGGAGTTC-3Ј c-Jun elevation is also seen in human neuropathies (Hutton et al., and 5ЈTGTCCCTCCAATTTTACACCTGTTCAATTC-3Ј (217bp band). 2011). Although c-Jun is implicated in the promotion of a num- The primers for the P0-Cre transgene were 5Ј-GCTGGCCCAAATGTT ber of tumors, in other situations, it may have a role in the pre- GCTGG-3Ј and 5ЈCCACCACCTCTCCATTGCAC-3Ј (480 bp band). vention of tumorigenesis by mechanisms that include activation of tumor suppressors such as P14 ARF/p19 ARFand Dmp1 (Eferl Nerve injury The sciatic nerve was exposed and crushed (3 ϫ 15 s at 3 rotation angles) and Wagner, 2003; Ameyar-Zazoua et al., 2005; Shaulian, 2010). ARF at the sciatic notch using angled forceps. The wound was closed using P19 is elevated in Schwann cells after nerve transection and veterinary autoclips. The nerve distal to the crush was excised for analysis the striking activation of Schwann cell c-Jun after injury is not at various time points. Contralateral uninjured sciatic nerves were used associated with tumor formation (Gomez-Sanchez et al., 2013; as controls for Western blotting, immunofluorescence, or electron Jessen et al., 2015b). Rather, the role of c-Jun is to take part in microscopy. controlling the conversion of myelin and Remak cells to Schwann cell specialized to perform injury-specific tasks and to promote Schwann cell culture repair (Jessen et al., 2015a; Jessen and Mirsky, 2016; Arthur- Schwann cell cultures were prepared from sciatic nerves of postnatal day 8 (P8)–P10 mouse pups essentially as in Morgan et al. (1991) and Arthur- Farraj et al., 2017; Gomez-Sanchez et al., 2017). This includes Farraj et al. (2011). After enzymatic dissociation and centrifugation, preventing the death of injured neurons and promoting axon the cell pellet was resuspended in defined medium (DM) (Meier et al., Ϫ6 growth by expression of trophic factors, guiding axons back to 1999) containing 10 M insulin and 5% horse serum (HS) and plated in their targets by forming regeneration tracks (Bungner bands), drops on coverslips coated with poly-L-lysine and laminin. Cells were and breakdown of myelin directly by autophagy and indirectly by incubated at 37°C/5% CO2 and allowed to adhere for 24 h. After 24 h, cytokine expression to recruit macrophages. Inactivation of the medium was changed to DM/0.5% HS (controls), DM with 10 Schwann cell c-Jun results in defective repair Schwann cells and ng/ml NRG1 alone, or DM with NRG1 10 ng/ml and dbcAMP 1 mM for impaired regeneration (Arthur-Farraj et al., 2012; Jessen and 48 h before fixation and immunolabeling. Mirsky, 2016). Antibodies The injury-induced extinction of myelin genes is also delayed The following antibodies were used for Immunofluorescence: c-Jun (Cell without c-Jun, indicating that c-Jun has a dual function, promot- Signaling Technology, rabbit 1:800, 9165, RRID: AB_2130165), Ki67 ing the expression of the repair phenotype and the suppression of (Abcam, rabbit 1:100, ab15580, RRID: AB_443209), Krox20 (Covance, the myelin phenotype. c-Jun suppression of myelin genes has rabbit 1:100, PRB-236P, RRID: AB_10064079), SOX-10 (R&D Systems, only been studied directly in culture, where c-Jun suppresses the goat 1:100, AF2864, RRID: AB_442208), donkey anti-goat IgG (HϩL) Krox20- or cAMP-induced activation of myelin genes, and en- Alexa Fluor 488 conjugate (Invitrogen, 1:1000, A11057, RRID: ϩ forced c-Jun inhibits myelination in cocultures (Parkinson et al., AB_2534104), Cy3 donkey anti-rabbit IgG (H L) (Jackson Immunore- 2004, 2008). Negative transcriptional regulation of myelination search, 1:500, 711-165-152, RRID: AB_2307443), biotinylated anti- rabbit IgG (GE Healthcare, 1:600, RPN1004, RRID: AB_1062582), and has also been shown for Notch1 and Sox2 in vivo and suggested Cy3 Streptavidin (Jackson ImunnoResearch, 1:500, 016-160-084). for other factors including Pax-3, Id2, and Sox-2 based on cell The following antibodies were used for Western blot: GAPDH (Sigma- culture experiments (Jessen and Mirsky, 2008; Roberts et al., Aldrich, rabbit 1:5000, G9545, RRID: AB_796208), calnexin (Enzo Life Sci- 2017). ences, rabbit 1:1000, ADI-SPA-860-D, RRID: AB_2038898), c-Jun (Cell The present results show that the function of c-Jun in Schwann Signaling Technology, rabbit 1:1000, 9165), Krox20 (Millipore, rabbit cells depends on gene dosage, and that Schwann cells are surpris- 1:500, ABE1374, RRID: AB_2715555), Mpz (Aves Labs, chick 1:2000, ingly tolerant of the moderately (ϳ6-fold) elevated c-Jun seen in PZO, RRID: AB_2313561), cyclin D1 (Santa Cruz Biotechnology, rabbit c-Jun OE/ϩ mice. In these mice, overexpression of c-Jun is suf- 1:200, sc-450, RRID: AB_627342), p19 Arf (5-c3-1) (Santa Cruz Biotech- nology, rat 1:100, sc-32748, RRID: AB_628071), anti-mouse IgG HRP- ficient to accelerate axonal regeneration (Wagstaff et al., 2017), so linked (Promega, 1:2000, W4028, RRID: AB_430834), anti-rabbit myelination and function are restored after nerve injury. Further, IgG, HRP-linked (Cell Signaling Technology, 1:2000, 7074, RRID: even high expression of c-Jun is not associated with tumor for- AB_2099233), anti-rat IgG, HRP-linked (Cell Signaling Technology, mation in Schwann cells, although this is sufficient to cause hy- 1:2000, 7077), and anti-chicken IgY, HRP-linked (Promega, 1:2000, pomyelination neuropathy. G1351, RRID: AB_430845). Fazal, Gomez-Sanchez et al. • Enforced Expression of c-Jun in Schwann Cells J. Neurosci., December 13, 2017 • 37(50):12297–12313 • 12299

Figure 1. Graded overexpression of c-Jun Schwann cell nuclei of c-Jun OE/ϩ and c-Jun OE/OE mice. A, Genomic structure of the c-Jun floxed allele in the Rosa26 locus. Excision of the stop codon iseffectedbycrossingRosa26c-Junff/ϩmicewithP0cre-expressingmicetogeneratec-JunOE/ϩandc-JunOE/OEmiceoverexpressingc-JunspecificallyinSchwanncells.B,PCRanalysisshowing the presence of c-Jun OE, Rosa26 WT, and P0 cre bands from DNA samples extracted from tails of WT, c-Jun OE/ϩ, and c-Jun OE/OE mice. C, Representative immunofluorescence images from WT, OE/ϩ, and OE/OE sciatic nerve cryosections showing Sox10- and c-Jun-positive nuclei. Note the graded increase in c-Jun in c-JunOE/ϩ and c-Jun OE/OE mice. Scale bar, 50 ␮m. D, Western blot of sciaticnerveproteinextractsfromP60WT,c-JunOE/ϩ,andc-JunOE/OEmiceshowingincreasingc-Junlevels.Thegraphquantifiesc-JunexpressioninWT(nϭ7),c-JunOE/ϩ(nϭ6),andc-Jun OE/OE (n ϭ 6) mice. The quantifications are normalized to the levels in uninjured WT nerves, which are set as 1. Note that the difference in c-Jun expression between c-Jun OE/ϩ and c-Jun OE/OE nerves is also significant. One-way ANOVA with Tukey’s comparison; *p Ͻ 0.05, ****p Ͻ 0.0001. E, Representative immunofluorescence images from purified Schwann cell cultures from WT and c-Jun OE/ϩ mice. The cells were exposed to neuregulin (nrg) alone or neuregulin plus cAMP analog (dbcAMP), a combination that mimics axonal myelination signals. Note that neuregulin plus dbcAMP suppresses c-Jun in WT, but not in c-Jun OE/ϩ, cells. Sox10 was used as a Schwann cell marker to show levels of c-Jun specifically in Schwann cells.

Immunohistochemistry RNA sequencing analysis Schwann cells were fixed in 4% paraformaldehyde/PBS for 15 min and Library preparation. Total RNA was isolated using the RNeasy Lipid then immunolabeled. Transverse sciatic nerve cryosections (10 ␮m) Tissue Mini Kit (Qiagen) with a DNase I step performed to eliminate were postfixed with 4% paraformaldehyde/PBS for 15 min, blocked in traces of genomic DNA.The purified mRNA was fragmented and primed 0.2% Triton X-100 with 10% HS in PBS and subsequently incubated with with random hexamers. Strand-specific first-strand cDNA was generated primary antibodies in blocking solution overnight at 4°C, followed by 2 h using reverse transcriptase in the presence of actinomycin D. The second in secondary antibodies and DAPI to identify cell nuclei (Thermo Fisher cDNA strand was synthesized using dUTP in place of dTTP to mark the Scientific, 1:50,000). For Ki67 staining, biotin antibodies followed by Cy3 second strand. The resultant cDNA was then “A-tailed” at the 3Ј end to streptavidin were used. prevent self-ligation and adapter dimerization. Truncated adaptors con- Western blotting taining a T overhang were ligated to the A-tailed cDNA. Successfully For blotting, homogenates were obtained from injured and uninjured ligated cDNA molecules were then enriched with limited cycle PCR nerves and from cultured nerve segments essentially as described previ- (10–14 cycles). The high-fidelity polymerase used in the PCR is unable to ously (Gomez-Sanchez et al., 2015). Experiments were repeated at least extend through uracil. This means that only the first strand is amplified three times with fresh samples and representative photographs are shown. for sequencing, thus making the library strand specific. Densitometric quantification was by Image Lab 4.1 (Bio-Rad Laboratories). Sequencing. Libraries to be multiplexed in the same run were pooled in Measurements were normalized to loading control GAPDH and/or equimolar quantities. Samples were sequenced on the NextSeq 500 in- calnexin. strument (Illumina), resulting in ϳ16 million reads per sample. 12300 • J. Neurosci., December 13, 2017 • 37(50):12297–12313 Fazal, Gomez-Sanchez et al. • Enforced Expression of c-Jun in Schwann Cells

Figure 2. Gene expression in c-Jun OE/ϩ and c-Jun OE/OE mice. A, Heat map of the 400 most regulated genes in uninjured nerves of WT and c-Jun OE/OE mice. B, Principal component analysis map of gene regulation in WT and c-Jun OE/OE nerves. C, Expression of 13 genes of interest in the sciatic nerve of OE/ϩ and OE/OE mice. The table shows how c-Jun elevation affects the expression ofasubsetofrepaircellmarkers,myelinproteins,andtranscriptionfactorsinthemouselinesindicated.Notethat,inOE/ϩnerves,onlyShhisregulatedՆ2-fold.InOE/OEnerves,repaircellmarkers are upregulated and myelin genes are downregulated, although two important myelin regulators, Krox20 and Sox10, are not strongly affected. WT, n ϭ 3; OE/ϩ, n ϭ 4; and OE/OE, n ϭ 4.

Data analysis. Run data were demultiplexed and converted to fastq files magnification. Higher magnifications were used to show the Remak bun- using Illumina’s bcl2fastq Conversion Software version 2.18 on BaseSpace. dles or any abnormal morphology found in the nerve. Fastq files were aligned to a reference genome using STAR on the BaseSpace The percentage of extracellular matrix was analyzed using ImageJ soft- RNA-Seq alignment app version 1.1.0. Reads per transcript were counted ware by converting electron microscopy images at 5000ϫ and 10,000ϫ using HTSeq and differential expression was estimated using the BioCon- into 8-bit images and tracing the presence of extracellular matrix. ductor package DESeq2 (BaseSpace app version 1.0.0). RNA Sequencing analysis was performed by UCL Genomics, UCL Behavioral tests Great Ormond Street Institute of Child Health. Experiments conformed to UK Home Office guidelines. Nine or more mice/genotype were tested. Six-week-old mice were tested before surgery Electron microscopy to ensure that there were no differences in normal responses between the Nerves were processed as described previously (Gomez-Sanchez et al., genetic backgrounds. Sciatic function index (Inserra et al., 1998), toe 2015). Transverse ultrathin sections from neonatal, P7, and P21 sciatic spread reflex (Ma et al., 2011), and the toe pinch test, modified from nerve or adult (P60) or aged (P300) sciatic nerves or from injured distal Collier et al. (1961), were performed as in Arthur-Farraj et al. (2012). stumps of adult sciatic nerves were taken 5 mm from the sciatic notch or cut site and mounted on film. Statistical analysis Photographs were taken using a Jeol 1010 electron microscope with a Results are expressed as mean Ϯ SEM. Statistical significance was esti- Gatan camera and software. Images were analyzed using ImageJ. Photo- mated by one-way ANOVA with Tukey’s correction, two-way ANOVA graphs at were taken at 3000ϫ to measure the number of myelinated with Bonferroni’s’s multiple comparisons, Mann–Whitney U test, or axons, non-myelinated axons bigger than 1.5 ␮m, and Schwann cell Student‘s t test. p Ͻ 0.05 was considered statistically significant. Statisti- nuclei. The nerve area was measured from photographs taken at 200ϫ cal analysis was performed using GraphPad Prism software (version 6.0). Fazal, Gomez-Sanchez et al. • Enforced Expression of c-Jun in Schwann Cells J. Neurosci., December 13, 2017 • 37(50):12297–12313 • 12301

Figure 3. The 15 most upregulated and downregulated genes in the sciatic nerve of OE/ϩ and OE/OE mice. A, B, The 15 most strongly elevated genes (top) and the 15 most suppressed genes (bottom) in response to c-Jun elevation in the mouse lines indicated. C, The 15 most strongly regulated gene in OE/OE nerves compared with expression in OE/ϩ nerves. D, Representative immunofluorescenceimagesfromWT,c-JunOE/ϩ,andc-JunOE/OEsciaticnervecryosectionsshowingKrox20inSox10-positiveSchwanncellnuclei.NotesimilarKrox20expressioninWTandc-Jun OE/ϩnerves,butmuchreducedlevelsinc-JunOE/OEnerves.Scalebar,50␮m.E,WesternblotofsciaticnerveproteinextractsfromP60miceshowingsimilarlevelsofKrox20inWTandc-JunOE/ϩ nerves,butlowerlevelsinc-JunOE/OEnerves.ThegraphquantifiesKrox20expressioninWT(nϭ5),c-JunOE/ϩ(nϭ4),andc-JunOE/OE(nϭ5)mice.Thequantificationsarenormalizedtothe levels in uninjured WT nerves, which are set as 1. One-way ANOVA with Tukey’s comparison: *p Ͻ 0.05, **p Ͻ 0.01. F, Western blot of sciatic nerve protein extracts from P60 mice. Note that Mpz expression is 15% lower than WT in c-Jun OE/ϩ nerves, but strongly suppressed in c-Jun OE/OE nerves. The graph quantifies Mpz expression in WT (n ϭ 5), c-Jun OE/ϩ (n ϭ 4), and c-Jun OE/OE (n ϭ 5) mice. The quantifications are normalized to the levels in uninjured WT nerves, which are set as 1. One-way ANOVA with Tukey’s comparison: ****p Ͻ 0.0001. 12302 • J. Neurosci., December 13, 2017 • 37(50):12297–12313 Fazal, Gomez-Sanchez et al. • Enforced Expression of c-Jun in Schwann Cells

Figure 4. Electron microscopic structure of adult nerves in WT and OE/ϩ mice. A, Electron micrographs showing similar overall appearance of nerves from P60 WT and OE/ϩ mice. Scale bar, 5 ␮m. B, Total area of P60 WT and OE/ϩ mouse nerves is not significantly different. Mann–Whitney U test: p ϭ 0.5317 (n ϭ 5). C, The number of Schwann cell nuclei per sciatic nerve profiles is not significantly different between P60 WT and OE/ϩ nerves. Mann–Whitney U test: p ϭ 0.0952 (n ϭ 5). D, Counts of Ki67-positive/Sox10-positive nuclei indicate that the difference in Schwann cell proliferation between WT and c-Jun OE/ϩ mice is not significantly different. Mann–Whitney U test: p ϭ 0.2000 (n ϭ 3). E, The total number of axons larger than 1.5 ␮m in diameter is similar in P60WTandOE/ϩnerves.Mann–WhitneyUtest:pϭ0.9444(nϭ5).F,Thepercentageofaxonsina1:1relationshipandϾ1.5␮mindiameterthataremyelinatedissimilarinP60WTandOE/ϩ nerves. Mann–Whitney U test: p Ͼ 0.9999 (n ϭ 5). G, Per nerve profile, the number of myelinated axons is similar in P60 WT and OE/ϩ nerves. Mann–Whitney U test: p ϭ 0.8016 (n ϭ 5). H, Per nerve profile, the number of axons in a 1:1 relationship and Ͼ1.5 ␮m in diameter but not myelinated is not significantly different between P60 WT and OE/ϩnerves. Mann–Whitney U test: p Ͼ 0.999 (n ϭ 5). I, Myelin thickness measured by g-ratios is thinner in P60 OE/ϩ nerves compared with WT. The whiskers extend from the 5th to the 95th percentiles. Mann–Whitney U test: p ϭ 0.0079 (n ϭ 5). J, The percentage of axons Ͼ1.5 ␮m in diameter that remain unmyelinated and within Remak bundles is very low and similar in P60 WT and OE/ϩ nerves. Mann–Whitney U test: p ϭ 0.1508 (n ϭ 5). K, Electron micrographs showing that the overall structure of adult P300 nerves in WT and OE/ϩmice is similar. Scale bar, 5 ␮m. L, The area of transverse profiles of P300 WT (n ϭ 4) and OE/ϩ (n ϭ 5) nerves is not statistically different. Mann–Whitney U test: p ϭ 0.2857. M, The number of Schwann cell nuclei per sciatic nerve profile is somewhat higher in P300 OE/ϩ (nϭ5)nervescomparedwithWT(nϭ4).Mann–WhitneyUtest:pϭ0.0317.N,Thetotalnumberofaxonslargerthan1.5␮mindiameterissimilarinP300WT(nϭ4)andc-JunOE/ϩ(nϭ5) nerves. Mann–Whitney U test: p ϭ 0.1905. O, The percentage of axons in a 1:1 relationship and Ͼ1.5 ␮m in diameter that are myelinated is similar in (Figure legend continues.) Fazal, Gomez-Sanchez et al. • Enforced Expression of c-Jun in Schwann Cells J. Neurosci., December 13, 2017 • 37(50):12297–12313 • 12303

Results Transcriptional profiling of uninjured nerves in WT, c-Jun Adult uninjured nerves of c-Jun OE/؉ and c-Jun OE/OE mice OE/؉, and c-Jun OE/OE mice have high levels of c-Jun protein in Schwann cell nuclei To document changes in gene expression caused by c-Jun el- A diagrammatic representation of how the c-Jun-overexpressing evation in c-Jun OE/ϩ and OE/OE mice, we performed RNA mice were bred and produced is shown in Figure 1A. The R26c- sequencing analysis on uninjured adult (P60) sciatic nerves. Junstopf mouse has a c-Jun cDNA insert in the Rosa26 WT locus Heat-map and principal component analysis confirmed that with two flanking loxP sites on either side of a STOP codon. These c-Jun overexpression was the dominant source of differential mice were bred with P0Cre ϩ/ Ϫ mice (Feltri et al., 1999). In the gene expression (Fig. 2A,B). In OE/ϩ nerves, which express ϳ6- presence of Cre recombinase, the STOP codon is removed and fold WT levels of c-Jun protein, 67 genes were Ն2-fold upregu- c-Jun is overexpressed specifically in Schwann cells (Feltri et al., lated and 25 genes were Ն2-fold downregulated compared with 1999). P0Cre Ϫ / Ϫ;R26c-Junstopf f/ϩ control mice will be referred WT nerves. Among 13 genes that we considered of particular to as WT mice, P0Cre ϩ/ Ϫ; R26c-Junstopf f/ϩ will be referred to as interest, one gene, Shh, was upregulated Ն2-fold (Fig. 2C). c-Jun c-Jun OE/ϩ mice, and P0Cre ϩ/ Ϫ; R26c-Junstopf f/f will be re- was expressed at 153% of WT levels and GDNF at 182% of WT ferred to as c-Jun OE/OE mice. Genotyping of WT, c-Jun OE/ϩ, levels and the myelin protein genes Mbp and Mpz were expressed and c-Jun OE/OE mice is shown in Figure 1B. at ϳ65% and 75% of WT levels, respectively. The mRNA level for We examined c-Jun protein expression in adult (P60) unin- Krox20 (Egr2), a key myelin regulator, was essentially unchanged. jured sciatic nerves of c-Jun OE/ϩ and c-Jun OE/OE mice and The 15 most upregulated and downregulated genes in c-Jun compared this with that seen in WT mice. Double immunola- OE/ϩ are shown in Figure 3A. beling with c-Jun antibodies and Sox10 antibodies to identify In OE/OE nerves, which express ϳ28-fold WT levels of c-Jun Schwann cell nuclei specifically showed that Schwann cells in protein, 909 genes were Ն2-fold upregulated and 1055 genes c-Jun OE/ϩ nerves expressed clearly elevated nuclear c-Jun levels were Ն2-fold downregulated compared with WT nerves. Most of compared with that seen in WT nerves, which showed barely the 13 genes of particular interest changed expression by Ն2-fold detectable c-Jun using this staining protocol (see Materials and in these mice (Fig. 2C).This included c-Jun, which was elevated 4- Methods). In c-Jun OE/OE nerves, nuclear c-Jun levels were fur- to 5-fold; Gdnf, which was elevated by ϳ56-fold; and Shh and ther increased (Fig. 1C). No increase in c-Jun was seen in Sox10- Olig1, which were elevated 20-fold and 48-fold, respectively. The negative nuclei labeled with DAPI (data not shown), indicating myelin protein genes Mbp and Mpz were reduced to 13–14% of that c-Jun overexpression in these mouse lines was Schwann cell WT levels. The 15 most upregulated and downregulated genes in specific, in agreement with previous observations (Feltri et al., c-Jun OE/OE nerves are shown in Figure 3B. 1999). A comparison of gene expression between OE/ϩ and OE/OE Western blotting showed that c-Jun protein levels in unin- mice with respect to the 13 genes of interest and the most regu- jured adult sciatic nerves were elevated ϳ6-fold in c-Jun OE/ϩ lated genes is shown in Figures 2C and 3C, respectively. mice and ϳ28-fold in c-Jun OE/OE mice compared with WT The fact that c-Jun protein was more strongly elevated in terms of (Fig. 1D). In c-Jun OE/ϩ mice, c-Jun mRNA levels were 4.5-fold fold change from WT than c-Jun mRNA in both c-Jun OE/ϩ and higher than in WT nerves. OE/OE mice suggests that posttranscriptional controls are impor- These data indicate that the axonal signals that normally sup- tant in controlling c-Jun levels. RNA-Seq data were deposited in press c-Jun during myelination in vivo fail to supress c-Jun ex- ArrayExpress (https://www.ebi.ac.uk/arrayexpress/) with accession pression from the c-Jun OE transgene, as expected (Jessen and ID: E-MTAB-6138. Mirsky, 2008; Parkinson et al., 2008). We verified this by expos- ing purified Schwann cell cultures to signals that mimic axonal myelin signals in mice, namely the combined activation of cAMP Expression of myelin-related proteins in uninjured nerves of ؉ and neuregulin pathways (Arthur-Farraj et al., 2011). In these OE/ and OE/OE mice experiments, a combination of 1 mM dbcAMP and 10 nM neu- We examined two key myelin-related proteins, the promyelin regulin failed to suppress nuclear c-Jun expression in c-Jun OE/ϩ transcription factor Krox20 (Egr2) and the myelin adhesion pro- Ϫ cells, although downregulation of c-Jun protein was seen in WT tein Pzero (Mpz), in uninjured sciatic nerves of c-Jun OE/ and cells (Fig. 1E). OE/OE mice. Consistent with the mRNA data, Krox20 levels were ϩ The elevation of c-Jun specifically in Schwann cell nuclei in essentially unaffected in c-Jun OE/ mice both in double-label c-Jun OE/ϩ and c-Jun OE/OE mice allowed us to study in vivo immunohistochemical experiments, which show Krox20 in the effects of a graded increase in c-Jun expression on Schwann Schwann cell nuclei, and in Western blotting experiments (Fig. ϳ cells in uninjured and injured nerves. 3D,E). Mpz levels in these mice were 15% lower than those found in WT mice (Fig. 3F). In contrast, the c-Jun OE/OE mice expressed significantly less Krox20 protein in Schwann cell nuclei 4 and Western blots (Fig. 3D,E) and much reduced levels of Mpz (Fig. 3F). (Figure legend continued.) P300 WT (n ϭ 4) and OE/ϩ (n ϭ 5) nerves. Mann–Whitney U test: This indicates that, in adult-Jun OE/ϩ nerves, the levels of key pϭ0.4444.P,ThenumbersofmyelinatedaxonspernerveprofileissimilarinP300WT(nϭ4) myelin-related proteins and their mRNA remain relatively mildly andOE/ϩ(nϭ5)nerves.Mann–WhitneyUtest:pϭ0.1905.Q,Pernerveprofile,thenumber affected despite ϳ6-fold elevation of Schwann cell c-Jun. This Ͼ ␮ of axons that are 1.5 m in diameter and in a 1:1 relationship but not myelinated is not tolerance does, however, break down when c-Jun levels are ele- significantly different between P300 WT (n ϭ 4) and OE/ϩ (n ϭ 5) nerves. Mann–Whitney U vated ϳ28-fold, as seen in c-Jun OE/OE mice, which is consistent test:pϭ0.4444.R,Measuredbyg-ratios,myelinisthinnerinP300OE/ϩ(nϭ5)nervesthan inWT(nϭ4).Thewhiskersextendfromthe5thtothe95thpercentiles.Mann–WhitneyUtest: with the capacity of c-Jun to regulate myelin genes negatively as p ϭ 0.0079. S, The percentage of unmyelinated axons Ͼ1.5 ␮m in diameter that remain in indicated in in vitro experiments and in mice with conditional RemakbundlesissimilarinP300WT(nϭ4)andOE/ϩ(nϭ5)nerves.Mann–WhitneyUtest: inactivation of Schwann cell c-Jun (Parkinson et al., 2004, 2008; p Ͼ 0.9999. Arthur-Farraj et al., 2012). 12304 • J. Neurosci., December 13, 2017 • 37(50):12297–12313 Fazal, Gomez-Sanchez et al. • Enforced Expression of c-Jun in Schwann Cells

Figure 5. High c-Jun levels in c-Jun OE/OE nerves result in hypomyelination. A, Electron micrographs showing lack of myelin and increased connective tissue spaces in P60 c-Jun OE/OE nerves comparedwithWT.B,Thetotalnumberofaxonslargerthan1.5␮mindiameterisnotsignificantlydifferentinP60WT(nϭ5)andc-JunOE/OE(nϭ3)nerves.Mann–WhitneyUtest:pϭ0.0714. C, The percentage of axons in a 1: 1 relationship and Ͼ1.5 ␮m in diameter that are myelinated is lower in c-Jun OE/OE mice than in WT (n ϭ 5) and OE/OE (n ϭ 3) nerves. Mann–Whitney U test: pϭ0.0179.D,Thenumberofmyelinatedaxonspernerveprofileissubstantiallyreducedinc-JunOE/OE(nϭ3)nervescomparedwithWT(nϭ5).Mann–WhitneyUtest:pϭ0.0357.E,Pernerve profile, the number of axons in a 1:1 relationship and Ͼ1.5 ␮m in diameter but not myelinated is much higher in OE/OE (n ϭ 3) nerves than in WT (n ϭ 5). Mann–Whitney U test: p ϭ 0.0179. F, Myelin, measured as g-ratios, is thinner in OE/OE (n ϭ 3) mice compared with WT (n ϭ 5). The whiskers extend from the 5th to the 95th percentiles. Mann–Whitney U test: p ϭ 0.0357. G, The percentageofunmyelinatedaxonsϾ1.5␮mindiameterthatremaininRemakbundlesishigherinOE/OE(nϭ3)nervesthaninWT(nϭ5)nerves.Mann–WhitneyUtest:pϭ0.0357.H,Nerves in c-Jun OE/OE mice (n ϭ 3) contain more mast cells that nerves in WT mice (n ϭ 5). Mann–Whitney U test: p ϭ 0.0179. I, OE/OE (n ϭ 3) nerves show more Schwann cell nuclei per nerve profile thanWT(nϭ5)nerves.Mann–WhitneyUtest:pϭ0.0357.J,WesternblotofsciaticnerveproteinextractsfromP60mice.NotethatlevelsofcyclinD1(amarkerofcellproliferation)aresignificantly higherinOE/OEnervesthaninWTorOE/ϩornerves.Theresultsarequantifiedinthegraph.WT,nϭ5;OE/ϩ,nϭ4;andOE/OE,nϭ5.Thequantificationsarenormalizedtothelevelsinuninjured WT nerves, which are set as 1. One-way ANOVA with Tukey’s comparison: *p Ͻ 0.05, **p Ͻ 0.01. K, Counts of Ki67-positive/Sox10-positive nuclei indicate (Figure legend continues.) Fazal, Gomez-Sanchez et al. • Enforced Expression of c-Jun in Schwann Cells J. Neurosci., December 13, 2017 • 37(50):12297–12313 • 12305

-Structure of adult nerves is nearly normal in c-Jun OE/؉ mice cates an impediment to myelination, a sorting defect was indi Although the levels of myelin proteins were normal in c-Jun cated by the fact that the percentage of Ͼ1.5 ␮m axons that OE/ϩ mice, it remained possible that the substantial c-Jun eleva- remained within Remak bundles was strikingly increased to tion affected myelination and nerve architecture. This was tested ϳ28% compared with Ͻ1% in WT (Fig. 5G). As a result of im- by a morphometric comparison of WT and c-Jun OE/ϩ nerves. paired myelination and sorting, the number of myelinated axons The general appearance of WT and c-Jun OE/ϩ nerves of in c-Jun OE/OE nerves was substantially lower than in WT nerves 60-d-old mice was similar (Fig. 4A). The size of the cross-sec- (Fig. 5D). As seen in mouse models of CMT1A neuropathy and a tional profiles of the sciatic nerve and the number of Schwann cell number of other mouse mutants with elevated, nontumorigenic nuclei were not significantly different between the two genotypes Schwann cell proliferation, the organization of Remak bundles and Ki67 labeling of Sox10-positive cells failed to show a signifi- was somewhat altered (Robertson et al., 2002; Chen et al., 2003; cant increase in Schwann cell proliferation (Fig. 4B–D). WT and Ling et al., 2005; Verhamme et al., 2011). The cells sometimes c-Jun OE/ϩ nerves had similar numbers of Ͼ1.5 ␮m axons per showed increased membranous structures and processes that were nerve profile and the percentage of segregated (1:1), myelin- not in contact with axons and they contained fewer axons per competent (Ͼ1.5 ␮m diameter) axons that were myelinated and transverse section of a bundle, suggesting the presence of a larger the total number of myelinated axons were comparable (Fig. 4E– number of Remak cells each taking care of fewer axons (Fig. 5A). G). Both WT and c-Jun OE/ϩ nerves contained similar, very low Increased mast cell numbers are seen in several neuropathic con- numbers of myelin-competent (Ͼ1.5 ␮m diameter) axons in a ditions and mutant models and after mechanical nerve injury 1:1 relationship that remained unmyelinated (Fig. 4H). Measure- (Olsson, 1971; Ling et al., 2005; Ishii et al., 2016). We therefore ments of g-ratios showed that myelin sheaths were slightly thin- counted mast cell numbers and found substantial elevation in ner in c-Jun OE/ϩ mice compared with WT (Fig. 4I). Remak c-Jun OE/OE nerves (Fig. 5H). bundles in c-Jun OE/ϩ mice appeared normal and the percent- An increase in Schwann cell number, a feature of many neu- age of Ͼ1.5 ␮m axons that were found within Remak bundles was ropathies including CMT1A (Robertson et al., 2002; Lupski and similar and very low in both genotypes (Fig. 4J). Chance, 2005), was also seen in c-Jun OE/OE nerves, the total We found that this similarity between WT and c-Jun-over- number of Schwann cell nuclei per nerve profile being ϳ6-fold expressing c-Jun OE/ϩ 60-d-old mice remained even in old (300 that in WT (Fig. 5I). Western blots of Cyclin D1 indicated ongo- d) mice (Fig. 4K–S). As in young mice, observations of general ing proliferation among the cells of c-Jun OE/OE nerves (Fig. 5J). appearance and quantitative analysis failed to reveal significant Proliferation of Schwann cells was indicated in double immunola- differences between the two genotypes, except for the difference beling with Ki67 and Sox10 antibodies to detect dividing Schwann in G-ratios, a difference that was also seen in young mice (Fig. cells because double labeled Schwann cells, although few, were 4R). The only age-induced change related to Schwann cell ϳ3ϫ more common in c-Jun OE/OE nerves than in WT nerves numbers, which were somewhat elevated in old mice, with the (Fig. 5K). Observations in the electron microscope provided no difference between the genotypes reaching statistical signifi- evidence for the presence of a significant number of Schwann cells cance (Fig. 4M). without contact with axons. The increased number of Schwann cell These observations show that, although c-Jun OE/ϩ mice nuclei in nerve sections is likely due to non-myelinating cells in a 1:1 show ϳ6-fold elevation of c-Jun protein that is localized to ratio, with axons being shorter than myelin cells, cells with thin my- Schwann cell nuclei, they achieve essentially normal Schwann cell elin sheaths being shorter than those with normal sheath thickness, and nerve architecture, with the exception of modestly reduced and an increased number of Remak cells. myelin thickness. The sciatic nerves of 60-d-old c-Jun OE/OE mice were enlarged, showing total cross-sectional profiles that were ϳ2ϫ that Adult nerves of c-Jun OE/OE mice are hypomyelinated and in c-Jun OE/ϩ or WT nerves (Fig. 5L). Collagen containing ex- show onion bulbs and hyperplasia but do not form tumors tracellular space was also markedly increased in c-Jun OE/OE In contrast to c-Jun OE/ϩ mice, the higher (ϳ28-fold) c-Jun nerves, occupying 133,349 ␮m 2 (Ϯ19,891; n ϭ 3; 54% of nerve expression in c-Jun OE/OE mice resulted in obvious lack of myelin area) in 60 d c-Jun OE/OE nerves, but only 16,069 ␮m 2 (Ϯ2834; in 60-d-old mice (Fig. 5A). Although the total number Ͼ1.5 ␮m n ϭ 5; 13% of nerve area) in WT nerves (Fig. 4M). This amounts axons was similar to WT (Fig. 5B), the percentage of segregated to an increase in extracellular space of 116,280 ␮m 2. Because the (1:1), myelin-competent (Ͼ1.5 ␮m diameter) axons that were nerves of c-Jun OE/OE nerves are 121,486 ␮m 2 larger than WT myelinated was reduced by ϳ40% (Fig. 5C) and there was a cor- nerves, Ͼ95% of the enlargement seen in c-Jun OE/OE nerves is responding increase in the number of myelin-competent axons due to increased collagen-containing extracellular space, with a that that had reached a 1:1 relationship but remained unmyeli- likely contribution from increased number of Remak cells and nated (Fig. 5E). The myelin sheaths in c-Jun OE/OE mice were cells other than Schwann cells. Increase in endoneurial connective also thin compared with WT (Fig. 5F). Although all of this indi- tissue is seen in a number of neuropathies, including CMT1A, and in the trembler and twitcher mouse mutants (Low, 1977; Palumbo et 4 al., 2002; Ling et al., 2005; Kagitani-Shimono et al., 2008; Fledrich et al., 2012). (Figure legend continued.) a higher rate of Schwann cell proliferation in c-Jun OE/OE mice (n ϭ We examined the mutant nerves extensively for the presence ϭ ϭ 5) compared with WT (n 3). Mann–Whitney U test: p 0.0179. L, The area of transverse of tumors or cellular arrangements reminiscent of tumor forma- ϭ ϭ profilesofOE/OE(n 3)nervesislargerthanofWTnerves.Mann–WhitneyUtest:p 0.0357 tion, but failed to find any evidence for either. Consistent with (n ϭ 5). M, Tracing of cell profiles and extracellular space (ECM) in transverse nerve sections, this, the tumor suppressor p19 ARF was strongly elevated in unin- followed by area measurements, shows a relative increase in extracellular space in OE/OE (n ϭ 3) nerves compared with WT (n ϭ 5). Mann–Whitney U test: p ϭ 0.0357. N, Western blot of jured nerves of c-Jun OE/EO mice (Fig. 5N). sciaticnerveproteinextractsfromP60mice.Noteincreasedexpressionofthetumorsuppressor Examination of old (P300) mice showed that only three of the p19ARF.Theresultsarequantifiedinthegraph.Thequantificationsarenormalizedtothelevels parameters studied above changed obviously with age (Fig. 6A–J): in uninjured WT nerves, which are set as 1. One-way ANOVA with Tukey’s comparison: *p Ͻ (1) the appearance of significant numbers of onion bulbs (Fig. 0.05 (n ϭ 3). 6F,G); (2) a reduction in Schwann cell proliferation, which was 12306 • J. Neurosci., December 13, 2017 • 37(50):12297–12313 Fazal, Gomez-Sanchez et al. • Enforced Expression of c-Jun in Schwann Cells

Figure 6. Nerves of aged c-Jun OE/OE mice. A, The nerves of P300 c-Jun OE/OE mice and WT mice contain comparable numbers of axons. Mann–Whitney U test: p ϭ 0.1143 (n ϭ 4). B, In P300 c-JunOE/OEmice,thepercentageofaxonsina1:1relationshipandϾ1.5␮mindiameterthataremyelinatedislowerthaninWTmice.Mann–WhitneyUtest:pϭ0.0286(nϭ4).C,InP300c-Jun OE/OE mice, the number of myelinated axons per nerve profile is reduced compared with WT. Mann–Whitney U test: p ϭ 0.0286 (n ϭ 4). D, Per nerve profile, P300 c-Jun OE/OE mice have a much larger number of unmyelinated axons that are Ͼ1.5 ␮m in diameter and in a 1:1 relationship and compared with WT mice. Mann–Whitney U test: p ϭ 0.0286 (n ϭ 4). E, The percentage of unmyelinated axons Ͼ1.5 ␮m in diameter that are found within Remak bundles is higher in OE/OE nerves than in WT nerves. Mann–Whitney U test: p ϭ 0.0286 (n ϭ 4). F, Electron micrographs from nerves of P300 c-Jun OE/OE mice showing examples of onion bulbs. The central axon, which is sometimes myelinated (top), is surrounded by relatively few layers of flattened Schwann cells, suggesting an early stage of bulb formation. Scale bar, 1 ␮m. G, The number of onion bulbs in P300 OE/OE nerves is much higher than in WT nerves. Mann–Whitney U test: p ϭ 0.0286 (n ϭ 4). H,P300OE/OEnervescontainahighernumberofmastcellsthanWTnerves.Mann–WhitneyUtest:pϭ0.0286(nϭ4).I,NervesinP300c-JunOE/OEmiceshowmoreSchwanncellnucleipernerve profile than nerves in WT mice. Mann–Whitney U test: p ϭ 0.0286 (n ϭ 4). J, The rate of Schwann cell proliferation is not significantly higher in P300 OE/OE nerves than in WT nerves, as shown by the counts of Ki67-positive/Sox10-positive nuclei. Mann–Whitney U test: p ϭ 0.1000 (n ϭ 3). no longer significantly elevated (Fig. 6J); and (3) a reduction in due to demyelination in the adult or inhibition of myelination the percentage of Ͼ1.5 ␮m axons that remained within Remak during development. bundles, from ϳ28% at P60 (Fig. 5G)toϽ2% (Fig. 6E). There- In developing nerves of c-Jun OE/ϩ mice, there was a trend fore, in nerves of c-Jun OE/OE mice, a large number of axons toward c-Jun elevation at P1, but this was not significant, whereas at appear to segregate gradually from Remak bundles between P60 P7, Jun was elevated ϳ6-fold compared with WT nerves at the same and P300. The proportion of these Ͼ1.5 ␮m axons that myelinate age. This failed to suppress levels of the myelin proteins Mpz and is similar to that of the Ͼ1.5 ␮m segregated axons in P60 nerves: Krox20 in Western blots (Fig. 7A,B). However, nuclear Krox20 was ϳ60% (Fig. 6B). No tumors were found in older mice (n ϭ 18). reduced, as shown by double labeling of nerve sections with Krox20 and Sox10 antibodies to identify Schwann cells (Fig. 7C). Developmental myelination is delayed in c-Jun OE/؉ mice In developing nerves of c-Jun OE/OE mice, c-Jun levels at P7 but inhibited in c-Jun OE/OE mice were ϳ8-fold that found in WT mice at that age and Mpz was sup- Although adult nerves of c-Jun OE/Ϫ mice are essentially nor- pressed in Western blots (Fig. 7A). Although Krox20 levels were not mal, we tested whether the c-Jun elevation in these nerves caused significantly reduced in Western blots, the number of Schwann cells a delay in myelination during development. We also determined that showed nuclear Krox20 was Ͻ50% of that in WT nerves, as seen whether the lack of myelin in the adult c-Jun OE/OE nerves was in double immunolabeling of nerve sections (Figure 7B,C). Fazal, Gomez-Sanchez et al. • Enforced Expression of c-Jun in Schwann Cells J. Neurosci., December 13, 2017 • 37(50):12297–12313 • 12307

Figure 7. Developmental overexpression of c-Jun delays myelination in c-Jun OE/ϩ mice, but inhibits myelination in c-Jun OE/OE mice. A, Western blot of nerve extracts from P1 and P7 sciatic nerves.Theresultsarequantifiedinthegraphs.DatafromP1nervesarenormalizedtolevelsinP1WTnerve,whicharesetas1,whereasdatafromP7nervesarenormalizedtolevelsinP7WTnerve, which are set as 1. Note that, by P7, c-Jun is elevated in both OE/ϩ and OE/OE nerves, whereas Mpz is reduced in OE/OE nerves only. P1 WT, n ϭ 3; OE/ϩ, n ϭ 3; P7 WT, n ϭ 4; OE/ϩ, n ϭ 4; and OE/OE, n ϭ 3. Statistical analysis for P1 is Student‘s t test: p ϭ 0.0608 for c-Jun, p ϭ 0.0174 for Mpz. Statistical analysis for P7 is one-way ANOVA with Tukey’s comparison: *p Ͻ 0.05, **p Ͻ 0.01, ****p Ͻ 0.0001. B, Western blot of nerve extracts from P7 WT, OE/ϩ, and OE/OE nerves. The results are quantified in the graph. Krox20 levels are similar in all genotypes. One-way ANOVA with Tukey’s comparison: p ϭ 0.2053 (n ϭ 3). C, The percentage of Krox20/Sox10-positive Schwann cells in sections from WT (n ϭ 8), OE/ϩ (n ϭ 6), and OE/OE (n ϭ 3) sciatic nerves at P7. Note a graded decrease in Krox20-positive cells as levels of c-Jun increase. One-way ANOVA with Tukey’s comparison: *p Ͻ 0.05, ***p Ͻ 0.001. D, Representative electron micrographs from P1, P7, and P21nervesofWT,c-JunOE/ϩ,andc-JunOE/OEmice.NotehypomyelinationinOE/OEnerveatP7andP21andtransienthypomyelinationinOE/ϩnervesatP7.Scalebar,5␮m.E,Thenerveareas aresimilarinallthreegenotypesatalldevelopmentalstages.One-wayANOVAwithTukey’scomparison:P1WT(nϭ5),OE/ϩ(nϭ4),andOE/OE(nϭ5),pϭ0.1978;P7WT(nϭ5),OE/ϩ(nϭ 5), and OE/OE (n ϭ 3), p ϭ 0.2261; and P21 WT (n ϭ 5), OE/ϩ (n ϭ 4), and OE/OE (n ϭ 4), p ϭ 0.084. F, The number of Schwann cell nuclei per sciatic nerve profile at P1, P7, and P21 in nerves of WT, c-Jun OE/ϩ, and c-Jun OE/OE mice. Note the transient difference between WT and OE/ϩ nerves at P7, whereas OE/OE nerves have more Schwann cells at P7 and P21. P1 WT, n ϭ 5; OE/ϩ, n ϭ 4; and OE/OE, n ϭ 5; P7 WT, n ϭ 5; OE/ϩ, n ϭ 5; and OE/OE, n ϭ 3; and P21 WT, n ϭ 5; OE/ϩ, n ϭ 4; and OE/OE, n ϭ 4. One-way ANOVA with (Figure legend continues.) 12308 • J. Neurosci., December 13, 2017 • 37(50):12297–12313 Fazal, Gomez-Sanchez et al. • Enforced Expression of c-Jun in Schwann Cells

Electron microscopy at P1, P7, and P21 showed that, in c-Jun adult P60 nerves. This includes a reduced number of myelinated OE/ϩ mice, myelination was transiently delayed at P7, whereas in axons, an increased number of segregated myelin competent axons c-Jun OE/OE mice, myelination was severely inhibited (Fig. 7D). that remained unmyelinated, thinner myelin sheaths, and an in- In c-Jun OE/ϩ mice, nerve area and the percentage of Ͼ1.5- creased number of Ͼ1.5-␮m-diameter axons that were seen within ␮m-diameter axons that were found within Schwann cell families Schwann cell families or Remak bundles (Fig. 6G–K). or Remak bundles, both of which were normal in the adult, were These experiments show that c-Jun negatively regulates devel- also normal during development at all 3 time points (Fig. 7E,K). opmental myelination in a dose-dependent manner. In the c-Jun However, a number of other parameters were abnormal at P7, OE/ϩ mouse, ϳ6-fold overexpression of c-Jun results in a tran- although they were normal at P1 and P21, revealing a transient sient delay at P7, whereas the nerve had recovered at P21. Con- delay in myelination. This includes the number of Schwann cell versely, in the developing nerves of c-Jun OE/OE mice, in which nuclei, which was elevated (Fig. 7F); the percentage of segregated c-Jun levels are ϳ50% higher than in c-Jun OE/ϩ nerves, myeli- (1:1), myelin-competent (Ͼ1.5 ␮m diameter) axons that were nation is permanently inhibited and seen in only 30–40% of myelinated, which was reduced; the total number of myelinated Ͼ1.5-␮m-diameter myelin-competent axons, a figure compara- axons, which was reduced (Fig. 7H); and the number of segre- ble to that seen in the adult. gated (1:1) myelin-competent (Ͼ1.5 ␮m diameter) axons that were not myelinated, which was elevated (Fig. 7I). In adult nerves Remyelination after nerve injury in c-Jun OE/؉ mice of these mice, the myelin sheaths were slightly thinner than in WT c-Jun is a key amplifier of the repair Schwann cell phenotype, and this difference was already present at P7 and P21 (Fig. 7J). which is generated in the distal stump of injured nerves. There- In the developing c-Jun OE/OE nerves, nerve area was not fore, elevation of c-Jun is a candidate approach for improving significantly different from that seen in WT or c-Jun OE/ϩ mice nerve repair under conditions where it falters, such as in older (Fig. 7E). The large nerve area in P60 nerves of these mice there- animals or due to long-term Schwann cell denervation (Wagstaff fore emerges in adulthood. At P7 and P21, these nerves contained et al., 2017). The observation that, in c-Jun OE/ϩ mice, adult ϳ2ϫ the number of Schwann cell nuclei seen in WT nerves, a nerves with ϳ6-fold elevation of c-Jun achieve a relatively normal smaller difference than that seen in the adult (Fig. 7F). This sug- degree of nerve architecture and myelination during develop- gests ongoing, low-level Schwann cell proliferation in adult ment, albeit with a delay, is encouraging for this approach mutant nerves, supported by Ki67 labeling of Sox10-positive because it demonstrates that significant c-Jun elevation and my- Schwann cells, although the differences between WT and c-Jun elination are compatible. However, after injury, remyelination is OE/OE nerves in cyclin D1 levels and did not reach significance slower and more easily disrupted than developmental myelina- (Fig. 7L,M). In other respects, the differences between developing tion. The two processes are also partly controlled by distinct sig- WT and mutant nerves at P7 and P21 already matched those seen in nals. We therefore tested the capacity of c-Jun OE/ϩ nerves to remyelinate after nerve injury. 4 After sciatic nerve crush injury, c-Jun levels distal to the crush were elevated in WT mice and this elevation was enhanced in (Figure legend continued.) Tukey’s comparison: **p Ͻ 0.01, ***p Ͻ 0.001, ****p Ͻ 0.0001. c-Jun OE/ϩ mice, as expected (Fig. 8A). At 1, 7, and 14 d after G,Thepercentageofaxonsina1:1relationshipandϾ1.5␮mindiameterthataremyelinated injury, c-Jun levels in OE/ϩ nerves were 2- to 3-fold higher than ϩ at P1, P7, and P21 in nerves of WT, c-Jun OE/ , and c-Jun OE/OE mice. Note reduced myelina- those in crushed WT control nerves. This amounted to an ϳ12- ϩ ϭ tion in OE/OE mice at P7 and P21 and transient reduction in OE/ mice at P7. P1 WT, n 5; fold (at 1 d after crush) to an ϳ30-fold (at 7 and 14 d after crush) OE/ϩ, n ϭ 4; and OE/OE, n ϭ 5; P7 WT, n ϭ 5; OE/ϩ, n ϭ 5; and OE/OE, n ϭ 3; and P21 WT, elevation of c-Jun in crushed c-Jun OE/ϩ nerves compared with n ϭ 5; OE/ϩ, n ϭ 4; and OE/OE, n ϭ 4;. One-way ANOVA with Tukey’s comparison: **p Ͻ 0.01and***pϽ0.001.H,ThenumberofmyelinatedaxonspernerveprofileatP1,P7,andP21 the levels found in uninjured control nerves. This was accompa- in nerves of WT, c-Jun OE/ϩ, and c-Jun OE/OE mice. Note the substantial reduction in myelin- nied by somewhat lower Krox20 levels (Fig. 8B). At 21 d after ated axons in OE/OE nerves at P7 and P21 and transient decrease in OE/ϩ mice at P7. P1 WT, crush, c-Jun levels in WT nerves had declined, although they n ϭ 5; OE/ϩ, n ϭ 4; and OE/OE, n ϭ 5; P7 WT, n ϭ 5; OE/ϩ, n ϭ 5; and OE/OE, n ϭ 3; and remained significantly above those in uninjured nerves (data not P21 WT, n ϭ 5; OE/ϩ, n ϭ 4; and OE/OE, n ϭ 4;. One-way ANOVA with Tukey’s comparison: shown). *p Ͻ 0.05, **p Ͻ 0.01, ***p Ͻ 0.001, ****p Ͻ 0.0001. I, The number of axons in a 1:1 When examined 4 d after nerve cut, nerves of c-Jun OE/ϩ relationship and Ͼ1.5 ␮m in diameter that remain unmyelinated at P1, P7, and P21 in nerves mice showed accelerated collapse/breakdown of myelin sheaths ϩ of WT, c-Jun OE/ , and c-Jun OE/OE mice. Note the increase in unmyelinated axons in OE/OE and faster clearance of the myelin protein MBP, in agreement ϩ ϭ ϩ ϭ ϭ nervesatP7andP21,butatP7onlyinOE/ mice.P1WT,n 5;OE/ ,n 4;andOE/OE,n with previous evidence that c-Jun promotes myelin clearance and 5;P7WT,nϭ5;OE/ϩ,nϭ5;andOE/OE,nϭ3;andP21WT,nϭ5;OE/ϩ,nϭ4;andOE/OE, myelin autophagy (Arthur-Farraj et al., 2012; Gomez-Sanchez et n ϭ 4. One-way ANOVA with Tukey’s comparison: **p Ͻ 0.01, ***p Ͻ 0.001, ****p Ͻ 0.0001. J, Reduction in myelin thickness, measured as g-ratios, which is seen in the adults in al., 2015)(Fig. 8C,D). ϩ bothOE/ϩandOE/OEnerves,isalreadypresentatP7andP21.P7WT,nϭ5;OE/ϩ,nϭ5;and Examination of crushed c-Jun OE/ nerves by electron mi- OE/OE, n ϭ 3; and P21 WT, n ϭ 5; OE/ϩ, n ϭ 4; and OE/OE, n ϭ 4. The whiskers extend from croscopy showed a significant delay in remyelination at 2 weeks the 5th to the 95th percentiles. One-way ANOVA with Tukey’s comparison: *p Ͻ 0.05, **p Ͻ after nerve crush (Fig. 8E). At this time point, ϳ35% of myelin- 0.01, ***p Ͻ 0.001. K, The percentage of unmyelinated axons that are Ͼ1.5 ␮m in diameter competent (Ͼ1,5 ␮m) axons were myelinated in OE/ϩ nerves, in Schwann cell families or Remak bundles at P1, P7, and P21 in nerves of WT, c-Jun OE/ϩ, and whereas Ͼ95% were myelinated in WT nerves (Fig. 8F). At 2 c-Jun OE/OE mice. Abnormally high numbers are seen in OE/OE nerves only. P1 WT, n ϭ 5; weeks after crush, the number of myelinated axons was also re- ϩ ϭ ϭ ϭ ϩ ϭ ϭ OE/ , n 4; and OE/OE, n 5; P7 WT, n 5; OE/ , n 5; and OE/OE, n 3; and P21 WT, duced in OE/ϩ nerves (Fig. 8G) and the number of segregated, ϭ ϩ ϭ ϭ Ͻ n 5;OE/ ,n 4;andOE/OE,n 4.One-wayANOVAwithTukey’scomparison:*p 0.05, myelin-competent axons without myelin was elevated (Fig. 8H). **p Ͻ 0.01, ***p Ͻ 0.001. L, Western blot of nerve extracts from P7 WT (n ϭ 3), OE/ϩ (n ϭ ϳ ϭ Significant recovery was seen 4 weeks after crush, when 75% of 3), and OE/OE (n 3) sciatic nerves showing cyclin D1, an indicator of cell proliferation. Quan- ϩ tificationofthedataarenormalizedtolevelsinP7WTnerve,whicharesetas1.CyclinD1levels myelin-competent axons were myelinated in OE/ nerves com- aresimilarinallmouselines.One-wayANOVAwithTukey’scomparison:pϭ0.3871.M,Counts pared with 98% in WT nerves. By 10 weeks, essentially all myelin- of Ki67-positive/Sox10-positive nuclei in P7 in WT (n ϭ 6), OE/ϩ (n ϭ 6), and OE/OE (n ϭ 3) competent axons were myelinated in both genotypes, a situation nerves. OE/OE nerves show increased Schwann cell proliferation. One-way ANOVA with Tukey’s similar to that in uninjured nerves of these mice (Fig. 8E–H). comparisons: *p ϭ 00121. In E–K, p-values are calculated relative to WT at the same age. Myelin sheaths in adult c-Jun OE/ϩ mice are thinner than in WT Fazal, Gomez-Sanchez et al. • Enforced Expression of c-Jun in Schwann Cells J. Neurosci., December 13, 2017 • 37(50):12297–12313 • 12309

Figure8. RemyelinationofOE/ϩnervesisdelayed. A,Westernblotofc-JuninnerveextractsfromthedistalstumpofadultWT(nϭ4)andOE/ϩ(nϭ4)nerves1d,7d,and2weeksaftercrush. The graph shows quantification of the results normalized to levels in uninjured WT nerves, which are set as 1. Note significant elevation of c-Jun at all time points. Mann–Whitney U test: 1 d, p ϭ 0.0006 (n ϭ 4); 7 d, p ϭ 0.0002 (n ϭ 4); and 2 weeks, p ϭ 0.0022 (n ϭ 4). B, Western blot of nerve extracts from the distal stump of adult WT and OE/ϩ nerves 2 weeks after crush. The results are quantified in the graph and normalized to levels in 2 week crushed WT nerve, which are set as 1. Krox20 levels are reduced in OE/ϩ nerves. Mann–Whitney U test: p ϭ 0.0286 (n ϭ 4). C, Representative electron micrographs from the distal stump 4 d after sciatic nerve cut in WT and c-Jun OE/ϩ mice illustrating collapsed myelin sheaths. The graph shows that fewer intact myelin sheaths per nerve profile remain in OE/ϩ nerves than in WT. Mann–Whitney U test: p ϭ 0.0286 (n ϭ 4). D, Transected c-Jun OE/ϩ nerves clear myelin protein faster than WT nerves. The graph shows the reduction in MBP 4 d after transection expressed as a percentage of MBP in uninjured nerve. WT and c-Jun OE/ϩ nerves have cleared close to 40% and 60% of their MBP content, respectively.ThedataareobtainedfromquantitationofWesternblots.WT,nϭ4;OE/ϩ,nϭ8.Mann–WhitneyUtest:pϭ0.0070.E,Representativeelectronmicrographsfromthedistalstump of WT and OE/ϩ nerves 2, 4, and 10 weeks after crush. In OE/ϩ nerves, the number of myelinated axons, which is reduced at 2 and 4 weeks, has recovered at 10 weeks. Scale bar, 5 ␮m. F, The percentageofaxonsϾ1.5␮mindiameterandina1:1ratiothataremyelinatedinthedistalstumpofWTandc-JunOE/ϩmice2,4,and10weeksafternervecrush.NotethatmyelinationinOE/ϩ nerves, which is reduced at 2 weeks, has recovered substantially by 4 weeks and is normal at 10 weeks. Mann–Whitney U test: 2 weeks, p ϭ 0.0286 (n ϭ 4); 4 weeks, p ϭ 0.0079 (n ϭ 5); and 10 weeks, p ϭ 0.0571 (n ϭ 4). G, The number of myelinated axons per nerve profile of the distal stump of WT and c-Jun OE/ϩ mice 2, 4, and 10 weeks (Figure legend continues.) 12310 • J. Neurosci., December 13, 2017 • 37(50):12297–12313 Fazal, Gomez-Sanchez et al. • Enforced Expression of c-Jun in Schwann Cells

(see previous section) and this difference was also seen in remy- c-Jun OE/OE mice confirms the potential of c-Jun to negatively elinated nerves (Fig. 8I). Tumors were not seen and regenerating regulate myelination, as seen previously in vitro. Myelination is WT and c-Jun OE/ϩ nerves did not differ in size (Fig. 8J) or other strongly impaired during development and this persists in adult aspects of general nerve architecture (Fig. 8E). nerves, which show hypomyelinating pathology, enlarged con- Importantly, the c-Jun OE/ϩ mice achieved full functional nective tissue, and immature onion bulbs. Fourth, even in nerves recovery after nerve crush. In the toe pinch test, which is primar- of aged c-Jun OE/OE mice, there is no evidence of tumor forma- ily a sensory test, time to full recovery was comparable in WT and tion and these nerves show strong activation of the tumor sup- c-Jun OE/ϩ mice, whereas time to initial response (group aver- pressor P19 ARF. The absence of tumorigenic effect of enforced age) was ϳ2 d longer in the mutants (Fig. 9A,B). In the toe spread c-Jun expression in Schwann cells is in agreement with the fact reflex, primarily a test of motor recovery, c-Jun OE/ϩ mice that mechanical nerve damage is not associated with tumor for- showed a transient delay in recovery on days 14 and 15 only (Fig. mation, although injured WT nerves contain proliferating cells 9C), possibly caused by delay in myelination at this time point with high c-Jun levels. (Fig. 8F,G). In the sciatic functional index, a sensory–motor test, c-Jun is involved directly or indirectly in the control ofϳ180 c-Jun OE/ϩ mice showed a nonsignificant trend toward a tran- of the ϳ4000 genes that change significantly after nerve injury. sient delay during the second and third week after injury (Fig. This allows c-Jun to take part in the regulation of a spectrum of 9D,E). properties of denervated repair Schwann cells, including morphology, autophagy-mediated myelin breakdown, and the Discussion expression of trophic factors linked to regeneration, including We have generated c-Jun OE/ϩ and c-Jun OE/OE mice with GDNF, artemin, BDNF, NGF, and LIF (Eggers et al. 2010; forced expression of c-Jun in Schwann cell nuclei to study the Arthur-Farraj et al., 2012; Fontana et al., 2012; Gomez-Sanchez et effects of a graded increase in c-Jun expression on Schwann cell al., 2015, Jessen and Mirsky, 2016; Jessen et al., 2015a). Of these, development and on remyelination after injury. First, we have GDNF, artemin, and LIF have been shown to be direct targets of shown that, during development and in adult nerves, Schwann c-Jun. Additional evidence for direct regulation of injury-induced cells are remarkably tolerant of elevated c-Jun levels. Although genes by c-Jun comes from a study of enhancer activation in developing and adult nerves of c-Jun OE/ϩ mice showed an ϳ6- Schwann cells. This showed c-Jun-binding sites associated with fold increase in c-Jun relative to WT nerves at the same age, injury-activated enhancers of genes elevated after nerve injury, myelination is only transiently affected at P7. By P21, myelina- including Shh, Olig1, and Runx2 (Hung et al., 2015). tion appears normal and, in the adult, Schwann cells and nerve The gene targets and function of AP-1 transcription factors, a architecture are similar to that in WT nerves, with the exception family to which c-Jun belongs, are regulated by dimerization part- of modestly reduced myelin thickness, which is unlikely to have ners and ancillary proteins (Chinenov and Kerrpola, 2001; Eferl significant consequences for sensory–motor control. Second, al- and Wagner, 2003). Little is known about these components in though remyelination after injury, which generally is more easily Schwann cells. perturbed than in development, is delayed in c-Jun OE/ϩ mice, RNA sequencing analysis showed that, in uninjured nerves of remyelination shows strong recovery at 4 weeks and essentially all c-Jun OE/ϩ mice, 95 genes were expressed at levels that differed myelin-competent axons are myelinated at 10 weeks after injury. Ն2-fold from those in WT nerves. Sixty-seven of these genes were As in uninjured nerves, the myelin sheaths of regenerated OE/ϩ upregulated in response to increased c-Jun levels including Shh. nerves remain thinner than those in regenerated WT nerves. The In c-Jun OE/OE nerves, 1964 genes were changed Ն2-fold and sensory and motor tests used here show only a slight delay fol- 909 of these were upregulated, among them GDNF, Shh, Olig1, lowed by complete functional recovery in c-Jun OE/ϩ mice. Id2, Sox2, and Runx2. The myelin genes Mpz, Mbp, and Pmp22 Therefore, the c-Jun elevation in c-Jun OE/ϩ mice is compatible were all strongly downregulated in c-Jun OE/OE nerves. Exam- with essentially normal restoration of myelin and nerve function ining injured nerves, we previously identified 172 genes that were similar to that found before injury. This is important because we expressed at different levels in 7 d cut nerves of mice, in which c-Jun find that the c-Jun elevation in c-Jun OE/ϩ mice is sufficient to was genetically inactivated compared with inured WT nerves. Of accelerate regeneration under conditions in which it is compro- these 172 genes, 106 were upregulated by higher c-Jun levels, mised by aging or long-term denervation (Wagstaff et al., 2017; namely expressed more highly in cut WT nerves than in cut c-Jun L.J. Wagstaff, J. Gomez-Sanchez, R. Mirsky, and K.R. Jessen, un- knock-out nerves. A comparison of the 15 genes most upregu- published data). Third, the higher overexpression achieved in lated by c-Jun in uninjured c-Jun OE/ϩ and OE/OE nerves in the present work with the 15 genes that are most upregulated by 4 c-Jun in 7 d cut nerves reveals only two common genes, Shh and (Figure legend continued.) after nerve crush. In c-Jun OE/ϩ mice, few myelinated axons are GDNF. This limited similarity indicates that the group of genes present at 2 weeks, but normal numbers are seen at 10 weeks. Mann–Whitney U test: 2 weeks, directly or indirectly regulated by c-Jun in Schwann cells that have p ϭ 0.0286 (n ϭ 4); 4 weeks, p ϭ 0.0079 (n ϭ 5); and 10 weeks, p ϭ 0.0571 (n ϭ 4). H, The adopted the repair phenotype after injury is significantly different number of unmyelinated axons Ͼ1.5 ␮m in diameter and in a 1:1 relationship that have not from the set of genes, which responds to c-Jun in cells that ensheath myelinated in the distal stump of WT and OE/ϩ nerves 2, 4, and 10 weeks after crush. Two and axons, many of which retain myelin differentiation. 4 week crushed OE/ϩ nerves contain elevated numbers of umyelinated axons, but their num- We find that the substantial c-Jun elevation in c-Jun OE/OE ϭ ϭ ber has fallen to normal levels at 10 weeks. Mann–Whitney U test: 2 weeks, p 0.0286 (n mice is sufficient to cause severe hypomyelination. This is un- ϭ ϭ ϭ ϭ 4); 4 weeks, p 0.0079 (n 5); and 10 weeks, p 0.1143 (n 4). I, Myelin thickness, doubtedly related to the ability of c-Jun to suppress myelin genes. measured as g-ratios, is reduced in the distal stump of OE/ϩ nerves at all time points after Although the causal relationship between the increased c-Jun crush. The whiskers extend from the 5th to the 95th percentiles. Mann–Whitney U test: 2 weeks,pϭ0.0286(nϭ4);4weeks,pϭ0.0079(nϭ5);and10weeks,pϭ0.0286(nϭ4). levels seen in human CMT1A and demyelination has not been J, The area of transverse sections through the distal stump 2, 4, and 10 weeks after crush is analyzed (Hutton et al., 2011), it seems clear that sustained dys- similarinWTandOE/ϩnerves.Mann–WhitneyUtest:2weeks,pϭ0.6571(nϭ4);4weeks, regulation of c-Jun resulting in high expression in uninjured p ϭ 0.3095 (n ϭ 5); and 10 weeks, p ϭ 0.9004 (n ϭ 4). In A and F–J, p-values are calculated nerves is a potential hazard. c-Jun is therefore a candidate for a relative to WT at the same time after injury. factor that could cause or promote pathological demyelination. Fazal, Gomez-Sanchez et al. • Enforced Expression of c-Jun in Schwann Cells J. Neurosci., December 13, 2017 • 37(50):12297–12313 • 12311

Figure9. Functionalrecoveryisslightlydelayedinc-JunOE/ϩmice.A,Toepinchassayshowingthepercentageofmicethatshowaresponsetoapinchoftoes3,4,and5atdifferenttimesafter sciatic nerve crush in WT and c-Jun OE/ϩ mice. In c-Jun OE/ϩ mice, all toes show a trend toward a delayed response. B, The average time in days after crush at which the first toe pinch response is seen in toes 3, 4, and 5. WT, n ϭ 10; OE/ϩ, n ϭ 9. Mann–Whitney U test: p-values are calculated relative to WT for each toe. Toe 3, p ϭ 0.0056; toe 4, p ϭ 0.0043; toe 5, p ϭ 0.0404. C, The toe spread reflex in WT (n ϭ 10) and OE/ϩ (n ϭ 9) mice after sciatic nerve crush. The reflex response is delayed at days 12, 14, and 15 in c-Jun OE/ϩ mice. Two-way ANOVA with Bonferroni’s comparison: p ϭ 0.0171. D, Representative digital footprints from WT and c-Jun OE/ϩ mice taken at 0, 7, 18, 21, 28, and 70 d after sciatic nerve crush used in sciatic functional index (SFI) analysis. E,SFIresultsfromWT(nϭ11)andc-JunOE/ϩ(nϭ8)miceatdifferenttimesaftersciaticnervecrush.ThereisnosignificantdifferencebetweenWTandc-JunOE/ϩmice.Two-wayANOVAwith Bonferroni’s comparison: p ϭ 0.5545.

c-Jun OE/OE nerves and nerves affected by demyelinating dles is 28% compared with Ͻ1% in WT. In common with human neuropathies, in particular CMT1A, show many similarities, most CMT1A nerves and nerves of the C22 and My41 mouse models of obviously hypomyelination. In c-Jun OE /OE mice, this involves CMT1A, c-Jun OE/OE nerves also contain increased Schwann substantially thinner myelin sheaths and an ϳ40% reduction in cell numbers (Robertson et al., 2002; Lupski and Chance, 2005). myelination among axons that have segregated and are myelin However, neither mouse models of CMT1A nor the c-Jun OE/OE competent. Axonal sorting is also adversely affected in younger mice show a significant numbers of Schwann cells that are without (P60) c-Jun OE/OE mice because, in these mice, the percentage of axonal contact. Increase in endoneurial connective tissue, which is unsorted Ͼ1.5-␮m-diameter axons that remain in Remak bun- substantial in c-Jun OE/OE mice, is also a feature of human CMT1A 12312 • J. Neurosci., December 13, 2017 • 37(50):12297–12313 Fazal, Gomez-Sanchez et al. • Enforced Expression of c-Jun in Schwann Cells nerves and of the CMT1A rat (Palumbo et al., 2002; Lupski and ome that define the Schwann cell repair phenotype after nerve injury. Cell Chance, 2005; Fledrich et al., 2012). Abnormalities of Remak Rep 20:2719–2734. CrossRef Medline fibers, including the formation of membranous structures that Chen S, Rio C, Ji RR, Dikkes P, Coggeshall RE, Woolf CJ, Corfas G (2003) Disruption of ErbB receptor signaling in adult non-myelinating Schwann do not contact axons, which are seen in c-Jun OE/OE mice, are cells causes progressive sensory loss. Nat Neurosci 6:1186–1193. CrossRef also described in the My41, C22, and C3 mouse models of Medline CMT1A (Robertson et al., 2002; Verhamme et al., 2011). Last, Chinenov Y, Kerppola TK (2001) Close encounters of many kinds: Fos-Jun onion bulbs, which are prominent in human CMT1A nerves and interactions that mediate transcription regulatory specificity. Oncogene seen in rodent CMT1A models (Lupski and Chance, 2005; Fle- 20:2438–2452. CrossRef Medline drich et al., 2012), are also present in nerves of aged c-Jun OE/OE Collier HO, Warner BT, Sker R (1961) Multiple toe-pinch method for test- mice. ing analgesic drugs. Br J Pharmacol Chemother 17:28–40. Medline De Felipe C, Hunt SP (1994) The differential control of c-jun expression in The histological changes outlined here for c-Jun OE/OE and regenerating sensory neurons and their associated glial cells. J Neurosci CMT1A nerves are generally not specific to these conditions, but 14:2911–2923. Medline are also observed to a varying degree in a number of other non- Eferl R, Wagner EF (2003) AP-1: a double-edged sword in tumorigenesis. tumor-associated human nerve pathologies, mutant mouse nerves, Nat Rev Cancer 3:859–868. CrossRef Medline or in injured nerves. (Low, 1977; Haney et al., 1999; Chen et al., Eggers R, Tannemaat MR, Ehlert EM, Verhaagen J (2010) A spatio-tem- 2003; Ling et al., 2005; Kagitani-Shimono et al., 2008; Ishii et al., poral analysis of motoneuron survival, axonal regeneration and neurotrophic factor expression after lumbar ventral root avulsion and 2016). The relative paucity of disease-specific structural changes implantation. Exp Neurol 223:207–220. CrossRef Medline in pathological nerves and the sloppy relationship between mo- Feltri ML, D’Antonio M, Previtali S, Fasolini M, Messing A, Wrabetz L lecular and histological phenotype makes it hard to interpret a (1999) P0-Cre transgenic mice for inactivation of adhesion molecules in particular histology in terms of a causal sequence. Schwann cells. Ann N Y Acad Sci 883:116–123. CrossRef Medline Although the availability of binding partners or other ancillary Fledrich R, Schlotter-Weigel B, Schnizer TJ, Wichert SP, Stassart RM, Meyer proteins are important regulators of c-Jun function, the levels of zu Ho¨rste G, Klink A, Weiss BG, Haag U, Walter MC, Rautenstrauss B, c-Jun protein are likely to be a key factor in determining whether Paulus W, Rossner MJ, Sereda MW (2012) A rat model of Charcot- Marie-Tooth disease 1A recapitulates disease variability and supplies bio- c-Jun has beneficial or adverse effects on nerve biology. Previous markers of axonal loss in patients. Brain 135:72–87. CrossRef Medline work shows that, already at low or moderate levels compatible Fontana X, Hristova M, Da Costa C, Patodia S, Thei L, Makwana M, Spencer- with myelination, c-Jun appears to promote neuron-supportive Dene B, Latouche M, Mirsky R, Jessen KR, Klein R, Raivich G, Behrens A signaling from Schwann cells to neurons, including the activation (2012) c-Jun in Schwann cells promotes axonal regeneration and motoneu- of trophic factors such as GDNF, whereas higher levels are ron survival via paracrine signaling. J Cell Biol 198:127–141. CrossRef required to suppress myelin genes. Therefore, in the C3 mouse Medline model of CMT1A, c-Jun is elevated, but this is not high enough Gomez-Sanchez JA, Gomis-Coloma C, Morenilla-Palao C, Peiro G, Serra E, Serrano M, Cabedo H (2013) Epigenetic induction of the Ink4a/Arf lo- to disrupt myelin, although it enhances axonal survival and sensory cus prevents Schwann cell overproliferation during nerve regeneration motor performance (Hantke et al., 2014). Similarly, in CMT1X and after tumorigenic challenge. Brain 136:2262–2278. CrossRef Medline mice, c-Jun is elevated and increases GDNF expression, but does Gomez-Sanchez JA, Carty L, Iruarrizaga-Lejarreta M, Palomo-Irigoyen M, not disrupt myelination (Klein et al., 2014). Varela-Rey M, Griffith M, Hantke J, Macias-Camara N, Azkargorta M, In sum, the present results show that, although moderate Aurrekoetxea I, De Juan VG, Jefferies HB, Aspichueta P, Elortza F, Aran- c-Jun increase is well tolerated during Schwann cell development say AM, Martínez-Chantar ML, Baas F, Mato JM, Mirsky R, Woodhoo A, Jessen KR (2015) Schwann cell autophagy, myelinophagy, initiates my- and remyelination after injury, strong elevation of c-Jun in elin clearance from injured nerves. J Cell Biol 210:153–168. CrossRef uninjured nerves suffices to induce significant hypomyelina- Medline tion pathology, implicating c-Jun in demyelinating neuropathies. Gomez-Sanchez JA, Pilch KS, van der Lans M, Fazal SV, Benito C, Wagstaff Conversely, we did not find a link between c-Jun elevation and LJ, Mirsky R, Jessen KR (2017) After nerve injury, lineage tracing shows tumorigenesis, which is consistent with the fact that tumors do that myelin and Remak Schwann cells elongate extensively and branch to not form after nerve injury, although c-Jun is strikingly elevated form repair Schwann cells, which shorten radically on remyelination. as Schwann cells lose myelin differentiation and proliferate. We J Neurosci. 37:9086–9099. CrossRef Medline ϩ Haney CA, Sahenk Z, Li C, Lemmon VP, Roder J, Trapp BD (1999) Hetero- also found that, after crush injury of OE/ nerves, myelination philic binding of L1 on unmyelinated sensory axons mediates schwann and nerve function can be restored in the presence of c-Jun levels cell adhesion and is required for axonal survival. J Cell Biol 146:1173– that are high enough to promote axonal regeneration in mice in 1184. CrossRef Medline which regeneration has been compromised by long-term dener- Hantke J, Carty L, Wagstaff LJ, Turmaine M, Wilton DK, Quintes S, Koltzen- vation or advanced age. burg M, Baas F, Mirsky R, Jessen KR (2014) c-Jun activation in Schwann cells protects against loss of sensory axons in inherited neuropathy. Brain 137:2922–2937. CrossRef Medline References Hung HA, Sun G, Keles S, Svaren J (2015) Dynamic regulation of Schwann Ameyar-Zazoua M, Wisniewska MB, Bakiri L, Wagner EF, Yaniv M, Weitz- cell enhancers after peripheral nerve injury. J Biol Chem 290:6937–6950. man JB (2005) AP-1 dimers regulate transcription of the p14/p19ARF CrossRef Medline tumor suppressor gene. Oncogene 24:2298–2306. CrossRef Medline Hutton EJ, Carty L, Laura´ M, Houlden H, Lunn MP, Brandner S, Mirsky R, Arthur-Farraj P, Wanek K, Hantke J, Davis CM, Jayakar A, Parkinson DB, Jessen K, Reilly MM (2011) c-Jun expression in human neuropathies: a Mirsky R, Jessen KR (2011) Mouse Schwann cells need both NRG1 and pilot study. J Peripher Nerv Syst 16:295–303. CrossRef Medline cyclic AMP to myelinate. Glia 59:720–733. CrossRef Medline Inserra MM, Bloch DA, Terris DJ (1998) Functional indices for sciatic, per- Arthur-Farraj PJ, Latouche M, Wilton DK, Quintes S, Chabrol E, Banerjee A, oneal, and posterior tibial nerve lesions in the mouse. Microsurgery 18: Woodhoo A, Jenkins B, Rahman M, Turmaine M, Wicher GK, Mitter R, 119–124. CrossRef Medline Greensmith L, Behrens A, Raivich G, Mirsky R, Jessen KR (2012) c-Jun Ishii A, Furusho M, Dupree JL, Bansal R (2016) Strength of ERK1/2 MAPK reprograms Schwann cells of injured nerves to generate a repair cell es- activation determines its effect on myelin and axonal integrity in the adult sential for regeneration. Neuron 75:633–647. CrossRef Medline CNS. J Neurosci 36:6471–6487. CrossRef Medline Arthur-Farraj PJ, Morgan CC, Adamowicz M, Gomez-Sanchez JA, Fazal SV, Jessen KR, Mirsky R (2008) Negative regulation of myelination: relevance Beucher A, Razzaghi B, Rhona Mirsky R, Jessen KR, Aitman TJ (2017) for development, injury, and demyelinating disease. Glia 56:1552–1565. Changes in the coding and non-coding transcriptome and DNA methyl- CrossRef Medline Fazal, Gomez-Sanchez et al. • Enforced Expression of c-Jun in Schwann Cells J. Neurosci., December 13, 2017 • 37(50):12297–12313 • 12313

Jessen KR, Mirsky R (2016) The repair Schwann cell and its function in Painter MW, Brosius Lutz A, Cheng YC, Latremoliere A, Duong K, Miller regenerating nerves. J Physiol 594:3521–3531. CrossRef Medline CM, Posada S, Cobos EJ, Zhang AX, Wagers AJ, Havton LA, Barres B, Jessen KR, Mirsky R, Arthur-Farraj P (2015a) The role of cell plasticity in Omura T, Woolf CJ (2014) Diminished Schwann cell repair responses tissue repair: adaptive cellular reprogramming. Dev Cell 34:613–620. CrossRef underlie age-associated impaired axonal regeneration. Neuron 83:331– Medline 343. CrossRef Medline Jessen KR, Mirsky R, Lloyd AC (2015b) Schwann cells: development and Palumbo C, Massa R, Panico MB, Di Muzio A, Sinibaldi P, Bernardi G, role in nerve repair. Cold Spring Harb Perspect Biol 7:a020487. CrossRef Modesti A (2002) Peripheral nerve extracellular matrix remodeling in Medline Charcot-Marie-Tooth type I disease. Acta Neuropathol 104:287–296. Kagitani-Shimono K, Mohri I, Yagi T, Taniike M, Suzuki K (2008) Periph- Medline eral neuropathy in the twitcher mouse: accumulation of extracellular ma- Parkinson DB, Bhaskaran A, Droggiti A, Dickinson S, D’Antonio M, Mirsky trix in the endoneurium and aberrant expression of ion channels. Acta R, Jessen KR (2004) Krox-20 inhibits Jun-NH2-terminal kinase/c-Jun Neuropathol 115:577–587. CrossRef Medline to control Schwann cell proliferation and death. J Cell Biol 164:385–394. Klein D, Groh J, Wettmarshausen J, Martini R (2014) Nonuniform molec- CrossRef Medline ular features of myelinating Schwann cells in models for CMT1: distinct Parkinson DB, Bhaskaran A, Arthur-Farraj P, Noon LA, Woodhoo A, Lloyd disease patterns are associated with NCAM and c-Jun upregulation. Glia AC, Feltri ML, Wrabetz L, Behrens A, Mirsky R, Jessen KR (2008) c-Jun 62:736–750. CrossRef Medline is a negative regulator of myelination. J Cell Biol 181:625–637. CrossRef Ling BC, Wu J, Miller SJ, Monk KR, Shamekh R, Rizvi TA, Decourten-Myers Medline G, Vogel KS, DeClue JE, Ratner N (2005) Role for the epidermal growth Roberts SL, Dun XP, Doddrell RDS, Mindos T, Drake LK, Onaitis MW, Florio factor receptor in neurofibromatosis-related peripheral nerve tumorigen- F, Quattrini A, Lloyd AC, D’Antonio M, Parkinson DB (2017) Sox2 esis. Cancer Cell 7:65–75. CrossRef Medline expression in Schwann cells inhibits myelination in vivo and induces Low PA (1977) The evolution of “onion bulbs” in the hereditary hypertro- influx of macrophages to the nerve. Development 144:3114–3125. CrossRef phic neuropathy of the Trembler mouse. Neuropathol Appl Neurobiol Medline 3:81–92. CrossRef Robertson AM, Perea J, McGuigan A, King RH, Muddle JR, Gabree¨ls-Festen Lupski JR, Chance PF (2005) Hereditary motor and sensory neuropathies AA, Thomas PK, Huxley C (2002) Comparison of a new pmp22 trans- involving altered dosage or mutation of PMP22: the CMT1A duplication and HNPP deletion. In: Peripheral neuropathy (Dyck PJ, Thomas PK, genic mouse line with other mouse models and human patients with eds), pp 1659–1680. Philadelphia: Elsevier. CMT1A. J Anat 200:377–390. CrossRef Medline Ma CH, Omura T, Cobos EJ, Latre´molie`re A, Ghasemlou N, Brenner GJ, van Shaulian E (2010) AP-1: the Jun proteins: oncogenes or tumor suppressors Veen E, Barrett L, Sawada T, Gao F, Coppola G, Gertler F, Costigan M, in disguise? Cell Signal 22:894–899. CrossRef Medline Geschwind D, Woolf CJ (2011) Accelerating axonal growth promotes Shy ME, Shi Y, Wrabetz L, Kamholz J, Scherer SS (1996) Axon-Schwann cell motor recovery after peripheral nerve injury in mice. J Clin Invest 121: interactions regulate the expression of c-jun in Schwann cells. J Neurosci 4332–4347. CrossRef Medline Res 43:511–525. CrossRef Medline Meier C, Parmantier E, Brennan A, Mirsky R, Jessen KR (1999) Developing Truett GE, Heeger P, Mynatt RL, Truett AA, Walker JA, Warman ML (2000) Schwann cells acquire the ability to survive without axons by establishing Preparation of PCR-Quality mouse genomic DNA with hot sodium hy- an autocrine circuit involving insulin-like growth factor, neurotrophin-3, droxide and tris (HotSHOT). BioTechniques 29:52–54. Medline and platelet-derived growth factor-BB. J Neurosci 19:3847–3859. Medline Verhamme C, King RH, ten Asbroek AL, Muddle JR, Nourallah M, Wolter- Morgan L, Jessen KR, Mirsky R (1991) The effects of cAMP on differentia- man R, Baas F, van Schaik IN (2011) Myelin and axon pathology in a tion of cultured Schwann cells: progression from an early phenotype long-term study of PMP22-overexpressing mice. J Neuropathol Exp Neu- (04ϩ) to a myelin phenotype (P0ϩ, GFAP-, N-CAM-, NGF-receptor-) rol 70:386–398. CrossRef Medline depends on growth inhibition. J Cell Biol 112:457–467. CrossRef Medline Wagstaff LJ, Gomez-Sanchez JA, Mirsky R, Jessen KR (2017) The relation- Olsson Y (1971) Mast cells in human peripheral nerve. Acta Neurol Scand ship between Schwann cell c-Jun and regeneration failures due to ageing 47:357–368. CrossRef Medline and long-term injury. Glia 65:E532.